Week 4 Assignment - Describe Threshhold, Ratio, Attack and Release
For this lesson, I want to take you though a compressor, what the main functions are, and show you how a simple change in the attack time can change how the way a compressor changes a track.
First, let's look at the compressor in Logic Pro, which is the DAW I like to use.
The window in the center shows the way three parameters will change the sound. Those three are the threshold, ratio, and knee. Threshold and ratio will have the greatest effect on your audio, but these three parameters are easier to show graphically, since they don't deal with time.
The threshold sets the level at which the compressor will start to work on the audio. Once the amplitude of the incoming audio is over the threshold, the compressor starts to reduce the level of the audio.
To show you how the threshold changes what the compressor does, here is what the compressor looks like with the threshold set to -8.5db.
Graphically, you can see how the threshold went up. The audio won't be changed by the compressor until it hits -8.5.
Now let's set the threshold to -35db.
The only thing we have changed is the threshold level, and look at how much of a difference that can make. Now the compression starts with the audio much lower, meaning that quieter passages will be lowered in volume as well.
The ratio tells the compressor how much it should lower the volume once the threshold is crossed. The numbers indicate how much amplitude comes in and how much will go out. In the above examples, the ratio is 3.7:1, so for every 3.7db of level coming in, the compressor will output 1db of the signal.
Let's change the ratio, and see how that affects the graph. This is what it looks like with a 2:1 ratio. The threshold is back to -20db.
We see how the slope of the compressor is barely there. For every 2db of signal, 1db is put out. This is pretty slight. Let's crank the ratio up to 23:1, with the same -20db threshold.
Now we are really squishing the audio. For every 23db of signal in, the compressor puts out 1db. We are still starting at that same -20db threshold, so that means we won't even come close to hitting a peak level and clipping.
The knee is a little harder to see, so I switched the compressor type to something that shows it a little better. Often a knee is either 'soft' or 'hard' with better compressors allowing you to change how much variation between the two extremes you will have. Knee adjustment is more subtle than the other setting in the compressor.
This is a very soft knee:
And this is a very hard knee:
The ratio is set to 3.7:1 and the threshold is -20db. Notice that the slope of the compression hasn't changed, just how we get in and out of that slope. One is more abrupt (hard knee) and one is more casual (soft knee). You might not hear the change most of the time, but when dealing with smoother audio (sweeps, crowd noise, etc), you might notice it.
Attack and Release times are the next on the list. Unlike threshold, ratio and knee, these are time based parameters. The others are amplitude based.
Attack tells the compressor how fast to start working once the threshold level has been reached. If I have the compressor set to 101ms (like the screenshot at the top), it will take the effects unit 101 milliseconds to go from not compressing to full compression.
Release tells the compressor how long to go back to unaffected operation after the audio level has gone below the threshold. Above, we have a 48ms release time. The compressor will take 48 milliseconds to stop compressing.
In class, they showed us how attack time can affect a drum track, but lets look at something else. Here is a voice track I worked on in Audacity.
I haven't processed this track yet, so let's open the compressor:
Everything here is pretty standard. I have a fast attack time, and for now, I'm not worried about the release time (here it is called 'decay'). After the process is done, the audio will be normalized, which means it will be evenly boosted to it's maximum volume without clipping. Let's see what that looks like
:
Some parts of the file got louder, while some got quieter. Look at midway point of the file, and how some of the audio is quieter. Were the release time shorter, the quieter parts there that fall below the threshold level wouldn't have been changed. But the compressor hadn't reached the end of it's release (decay) time and was therefore still compressing.
Let's make one small change to the settings. Let's bring the attack time up to 0.2 seconds.
Now look at what happens:
Many parts, most dramatically the first few words, have a large initial volume, but then get much quieter. The first bit of the audio isn't affected, because the compressor takes longer to react after the threshold has been crossed.
Here are all three tracks next to each other, so you can see how they compare:
And here are the three tracks, back to back for you to listen to.
The difference between the three tracks is actually pretty stunning. All for a simple attack time change.
Keep in mind that Audacity has a lot fewer options than most DAWs. It's big selling point is it's free. I use Logic Pro for a reason. :-)
Thanks for going though this tutorial. I hope it was helpful to you.
